skills.rest
microsoft
microsoftOFFICIAL
Explore all skills by this author →
Website𝕏@OpenAtMicrosoftGitHub

rpi-challenger

Challenges confirmed tasks, plans, and artifacts through adaptive skeptical questions to expose assumptions.

1.4k|284|Updated Nov 2, 2025
Search Tags:#decision analysis#plan validation#assumption challenging#gitHub copilot#skeptical questioning#rpi workflow
One-click install
npx skills add https://github.com/microsoft/hve-core --skill rpi-challenger
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rpi-challenger
Source: https://github.com/microsoft/hve-core/tree/main/.github/skills/rpi/rpi-challenger
Command: npx skills add https://github.com/microsoft/hve-core --skill rpi-challenger

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Teams often act on tasks, decisions, or plans that contain hidden assumptions, unclear boundaries, or missing evidence. This Skill examines a confirmed subject through adaptive, skeptical questioning so material uncertainties surface before downstream work begins.

Core Features & Use Cases

  • Adaptive Challenge Questioning: Asks one focused, open-ended, non-leading question per turn, choosing angles based on evidence and prior answers rather than a fixed checklist.
  • Durable Session Records: Creates and maintains a challenge record at .copilot-tracking/challenges/ capturing confirmed scope, evidence basis, Q&A log, coverage, and unresolved items.
  • Read-Only Boundary: Treats product and source artifacts as read-only, writing only the challenge session record and never invoking other agents or RPI stages.
  • Use Case: Before implementing a plan, run the challenger on the plan document to expose unvalidated assumptions, then route unresolved items to research, planning, or review stages.

Quick Start

Ask the assistant to challenge a confirmed plan or decision, for example: challenge the migration plan in docs/plan.md with focus on data-loss risks.

Frequently Asked Questions about rpi-challenger

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I challenge a plan or decision before implementing it?▼

Invoke the rpi-challenger skill with the subject and optional artifacts or focus, for example subject=plan artifacts=docs/plan.md. It confirms the scope with you, then asks one open-ended question per turn to surface assumptions and records everything in a challenge session file.

What is the RPI challenger skill in hve-core?▼

It is a GitHub Copilot skill in the hve-core repository that examines confirmed tasks, decisions, plans, or artifacts through skeptical questioning. It is part of the RPI (research, plan, implement) workflow and produces a durable challenge record with coverage and unresolved items.

Does the challenger modify my source files or plans?▼

No. The skill enforces a strict read-only boundary over product and source artifacts. Its only writable output is the challenge session record created under .copilot-tracking/challenges/ in your workspace.

When should I not use adversarial questioning on a task?▼

Avoid it when the subject is not yet confirmed or lacks enough context to form a factual scope, since the skill will stop and ask for the smallest missing artifact or fact. It also does not replace review or validation, as it never judges correctness or recommends solutions.

What happens to unresolved assumptions after a challenge session?▼

Unresolved items are recorded with the smallest missing evidence or decision needed to resolve them. The skill then advises an exact next /rpi-research, /rpi-plan, /rpi-implement, or /rpi-review command only when an unresolved item warrants it, without invoking it.

Related Skills

Explore More Software Engineering Skills→
foyzulkarimfoyzulkarim
41

start-task

Detect task identifiers, sync the main branch, and create a named branch with saved context.

Community
Advanced
tincoppertincopper
10

grill-me

Interrogate plans and designs by questioning assumptions, dependencies, and decisions.

Community
Intermediate
J03Fr0stJ03Fr0st

grill-me

Conduct a structured question-by-question interview to challenge and refine assumptions.

Community
Intermediate
dmooneydmooney

rundale-bench

Evaluate Rundale model quality with Sonnet 4.6 rubric scoring.

Community
Advanced